Journal Archive

Fig. 1 shows the location of two sampling sites and the meteorological station in Dhaka, where long term PM samples have been collected over past decade. The hot spot site in Dhaka (HS) is a high traffic area located at Farm Gate adjacent to the intersection of several main roads typically carrying high volumes of traffic (Latitude: 23.76°N; Longitude: 90.39°E). The semiresidential area (SR) (Latitude: 23.77°N; Longitude 90.38°E) in Dhaka is located in the Atomic Energy Centre, Dhaka (AECD) campus building with moderate traffic. Long term sampling (April 2000 to March 2006 at HS site and April 2000 to March 2010 at SR site) at both the sites were performed using a ‘Gent’ stacked filter sampler (Hopke et al., 1997) capable of collecting air particulate samples in coarse (2.2 to 10 μm) and fine (2.2 μm) size fractions. The sampling protocol was described by Begum et al. (2011a).

Fig. 1. 
Map of Dhaka, Bangladesh showing the locations of the two sampling sites (HS site and SR site).

For the ultrafine particulate sampling campaign, two MOUDI and three AirMetrics MiniVol samplers were used. The samplers were placed on the flat roof of the guardhouse of the Bangladesh Agricultural Research Council (BARC) at Farm Gate.

The10-stage micro-orifice uniform deposit impactors (MOUDIs, MSP Corp., Model 100) were operated in parallel for consecutive 24 hour periods from 4 January to 13 January 2010 to collect a total of 10 samples. The cut points for the 10 stages are 18, 10, 5.6, 3.2, 1.8, 1.0, 0.56, 0.32, 0.18, 0.1 and 0.056-μm. However, an AIHL-design cyclone separator (John and Reischl, 1980) was connected upstream to each MOUDI to remove large particles having a 50% aerodynamic cutoff diameter at 1.8 μm before the air passed through the MOUDI impactors at a flowrate of 30 L/min. The cyclone was used to prevent particle bounce in the upper stages of the MOUDI since our interest in the study was to quantify mostly PM_1.0 and smaller size ranges. The AIHL cyclone was used to suppress particle bounce from the upper stages of the impactors. Particles in the aerodynamic diameter size range of 0.056 to 1.8 μm were collected in impaction stages 5 through 10 of the impactors. All 10 stages of the impactors were in place during the experiment but only the lower 6 stages collected meaningful samples because of the cyclone separator upstream of the impactors. The flow rate, and start and end time were recorded to allow the total volumes and mass concentrations to be calculated.

One impactor was loaded with aluminum foil substrates (MSP Corp., 47 mm diameter) and a 47 mm diameter quartz fiber backup filter. The second impactor was operated with PTFE impaction substrates and a backup filter (47 mm diameter, Teflo filters (Pall Corp.), 2.0 μm pore size). The foil impaction substrates and quartz fiber filters were baked before use in order to lower their carbon blank values, since these materials were dedicated to the analysis of carbonaceous aerosol species; foil substrates were baked for 48 hours at 500°C, and quartz fiber filters were baked for 12 hours at 500°C. To avoid contamination by organic compounds, no grease or oil was applied to the impaction substrates. Following sample collection, filter samples and samples on impaction substrates were placed in petri dishes, and then frozen immediately until subsequent analysis.

Two AirMetrics MiniVol samplers were used to collect PM₁₀ and PM_2.5 on Teflon filters and a third Mini- Vol sampler was used to collect PM_2.5 on quartz filters for subsequent compositional analysis parallel to the MOUDI sampling.

The masses of the samples were determined by weighing the Teflon filters before and after their exposure. A ²¹⁰Po (alpha emitter) electrostatic charge eliminator (STATICMASTER) was used to eliminate the static charge accumulated on the filters before each weighing. Elemental analyses were performed on the Teflon filters using PIXE (Cohen et al., 1996), and XRF methods in GNS Science, New Zealand and Clarkson University, NY, USA, respectively. These methods are described in detail by Landsberger and Creatchmam (1998).

Carbonaceous materials collected on the PM_2.5 quartz filters were analyzed via IMPROVE/TOR protocol (Chow et al., 2008, 2004, 1993) for eight temperature resolved carbon fractions (Atmoslytic model DRI2100A). This protocol provides concentrations of carbon in four organic carbon fractions (OC1, OC2, OC3, OC3), three elemental carbon fractions (EC1, EC2, EC3) and pyrolyzed carbon (OP). The samples which were collected on Teflon filters in PM_2.5 were analyzed by an EEL-type Smoke Stain Reflectometer (Diffusion Systems, Ltd) to measure BC. Secondary standards of known black carbon concentrations were used to calibrate the reflectometer (Biswas et al., 2003). The aluminum foil substrates (MOUDI sampler) were analyzed via IMPROVE/TOR protocol (Chow et al., 2008, 2004, 1993) for eight temperature resolved carbon fractions.

In Bangladesh, the climate is characterized by high temperatures and high humidity for most of the year, and distinctly marked seasonal variations in precipitation. Based on the meteorology, the year can be divided into four seasons, pre-monsoon (March-May), monsoon (June-September), post-monsoon (October-November) and winter (December-February) (Salam et al., 2003). The winter season is characterized by dry soil conditions, lower relative humidity, scanty rainfall and low northwesterly winds. The meteorological data used in this study was obtained from a local meteorological station that is located about 2 and 5 kilometers north of the hot spot (HS) and semi-residential (SR) sites, respectively.

Dhaka has a sub-tropical climate and in under the influence of the strong southwest or summer monsoon and weak northeast or winter monsoon. It has been understood from last few years of air quality monitoring, the air quality level of the city greatly influenced by the Asian monsoon. The air quality characteristics in Dhaka shows distinct seasonal variations, with high pollution episodes observed during the winter, while summer has relatively cleaner ambient air. During the dry winter and part of the post-monsoon season, the strength of north and northwesterly winds coming from India, Nepal and Southeastern China to the Bay of Bengal through Bangladesh can transport pollutants to the city. Moreover, during this season, wind speeds are low so that locally emitted pollutants are not well dispersed into the downwind area. Mixing heights of the boundary layer are important factors influencing the dispersion of pollutants.

Samples were collected by three AirMetrics MiniVol samplers and the two MOUDI samplers. During the study period, the daily PM_2.5 and PM₁₀ values (Table 1) were higher than the 24 h average Bangladesh National Ambient Air Quality Standard (BNAAQS), which were set at 65 μg/m³ and 150 μg/m³, respectively. The ratios of EC/PM_2.5 concentrations for the intensive study and for the same period in previous years were calculated and compared (Fig. 2). It was observed that the ratios have decreased after 2002. Although there is evidence that BC comes from the transboundary sources, especially during winter (Begum et al., 2011a, b, 2008; Ramanathan and Carmichael, 2008), it has significant local sources. BC concentrations were observed to vary from month to month depending on the meteorological conditions. It has been found that the concentration has decreased relative to previous years as a result of policies adopted by the government (Begum, 2006a).

Fig. 2. 
Variation of EC/PM_2.5 ratio in different years from 1 to 14 January.

PM and BC concentrations in the AirMetrics and MOUDI samples were compared. Fig. 3 shows mass ratios for different fractions of PM during the period of 4 January to 13 January 2010. From the average mass ratios of different PM fractions, it was observed that about 78% of PM₁₀ is PM_2.5 and about 63% of PM_2.5 is PM₁. It can be seen that the BC and EC masses are concentrated mainly in the PM₁ fractions (Fig. 3). Thus, most of the BC mass was in the accumulation mode (0.1-1 μm diameter). Prior studies (Begum et al., 2006a, b; Salam, 2003) suggested that high EC contributions come mainly from the diesel trucks, petrol vehicle, two-stroke motorcycles, diesel-powered generators, open burning incinerators, and brick kilns.

Fig. 3. 
Mass ratios for different fractions of PM during the January 2010.

Black carbon (BC) and elemental carbon (EC) in airborne particulate matter originate from the incomplete combustion of carbonaceous fuel. These terms are used to describe aspects of ambient particulate matter. However, they are not interchangeable because they are each defined by the way they are measured. Salako et al. (2011) report the variability of EC and BC in a number of locations across Asia and the South Pacific region. It was found that the BC concentrations were comparable with the EC values (slope=1.1, r²=0.91) (Fig. 4). The mass absorption coefficient that is used for BC calculation was 10.98m²/g (Biswas et al., 2003). The two major classes of sources for EC and OC, fossil fuels and biomass based fuels exhibit distinctly different EC/TC and K/EC ratios allowing for source contributions to be estimated for EC and OC in the aerosol. Emissions from diesel engines as well as from oil- and coal-fired stationary sources exhibit EC/TC ratios in the range of 0.6-0.7 (Rau, 1989; Williams et al., 1989; Cass et al., 1982). It was found that K/EC ratio is 0.1 (Salam et al., 2003) for PM_2.5, that indicate little influence from biofuel sources (Table 2). In smoke from biomass fires, the K/BC ratio is characteristically higher than for the other sources of EC. This K/EC ratio (rural sites) varies from 0.22 to 1.4 depending on the kitchen environment and the fuel types (Samir, 2007; Venkaraman et al., 2002; Maenhaut et al., 1996). Hence, it can be concluded that biomass fuel combustion is not a major source of EC in Dhaka urban atmosphere.

Fig. 4. 
Relation between EC and BC during January 2010.

EC has different sources such as diesel engine emissions, fossil fuel combustion at low burning efficiency and biomass burning. Figs. 5 and 6 show the fraction of total carbon contributed by each of the six carbon fractions for each sample (EC3=0) collected with the AirMetrics and MOUDI samplers, respectively. The abundance of EC1 is generally higher in gasoline vehicles than in diesel vehicles (Watson et al., 1994). Similarly the OC1 abundance is higher in the diesel fueled relative to gasoline fueled and OC2 would be higher in gasoline emissions compared to diesel. In the present study, it was found that the contribution of gasoline is much higher than diesel. This finding could be due to banning of heavy duty diesel trucks in Dhaka city during daytime and use of gasoline during starting of CNG converted engines. The contribution of EC from compressed natural gas (CNG) would be much less.

Fig. 5. 
Fractions of OC1, OC2, OC3, OC4, EC1, and EC2 in PM_2.5 mass during the January 2010.

Fig. 6. 
Fractions of OC1, OC2, OC3, OC4, EC1, and EC2 in PM₁ mass during January 2010.

The population of motorized vehicles in Dhaka is more than 560,000 up to 2010 and average annual growth is about 10% (Fig. 7). Motorcycles, which are mainly gasoline fueled vehicles, are the largest fleet and comprise about 35% of the total vehicle population. Motor cars and jeep/station wagon/microbus are the next largest fleet component (about 34%). Because of government policy, a large number of vehicles especially cars, taxis, and three wheelers are now operated on CNG (as of January 2011 over 0.12 million vehicles were converted to CNG). The remaining vehicles are using either diesel or gasoline (BRTA, 2010). Heavy duty commercial vehicles comprising 19% of the total fleet are expected to be causing air pollution.

Fig. 7. 
Year-by-year registration of vehicles in Dhaka City (Source: Bangladesh Road Transport Authority, 2010).

Long-term PM data (April 2000 to March 2006) and (April 2000 to March 2010) in two size fractions (PM_2.2 and PM_2.2-10) obtained from two particulate air quality-monitoring stations, (HS and SR) were also analyzed to see the trend of air quality in Dhaka. The HS site is a heavily traffic-impacted site while the SR site is in a residential area. The sources of the observed PM for these sites have been extensively discussed in prior source apportionment studies (Begum et al., 2010, 2008, 2006a, b, 2004).

The year-by-year average values of PM₁₀, PM_2.2, and BC at mass concentrations obtained in these sites are presented in Fig. 8. From the figure, it can be observed that over the years, the PM_2.2 mass concentration has decreasing trend while PM₁₀ has increased, which suggests that PM_2.2-10 fraction has been increasing. The decrease in fine particle mass is notable given the rise in the number of vehicles noted in Fig. 7. Thus, the policy approaches to reduce vehicular emissions have provided some of the improvements in the observed fine PM mass values.

Fig. 8. 
The yearly average PM₁₀, PM_2.2 and BC mass concentrations from April 2000 to March 2006 and April 2000 to March 2010 at HS and SR sites respectively.

The reduction in PM_2.2 fraction of the particulate pollution seems to be over compensated by increasing the coarse (PM_10-2.2) fraction and major portion of those particles are of natural origin. Increases in soil and road dust particles in the coarse fraction increase the PM₁₀ mass concentration. About 60 to 70% of coarse mass is soil dust including road dust (Begum, 2011a) in the Dhaka air. BC concentrations also show a decreasing trend. Further policy interventions may also be require to reduce such coarse particles, which are mainly originated from soil, road dust and construction activities.

The yearly average PM_2.2/PM₁₀ and BC/PM_2.2 ratios are also shown in Fig. 9. It may be seen that after 2001/2003 the average BC/PM_2.2 ratio remained unchanged although major sources such as vehicles have significantly increased. During this period, different policies were implemented to reduce the ambient PM concentrations. These measures included a phased reduction of gasoline-powered, two-stroke, three-wheel taxis with their complete banning as of January 2003 (Begum et al., 2006b), improved training of engine mechanics, and removal of trucks and buses that were more than 20 years old, and also electronic traffic signals.

Fig. 9. 
Ratios of PM_2.2/PM₁₀ and BC/PM_2.2 over the collection periods from April 2000 to March 2006 and April 2000 to March 2010 at HS and SR sites respectively.

It can be seen that the mean PM_2.5 values were much higher than the 2006 USEPA PM National Ambient Air Quality Standards as well as the Bangladesh National Ambient Air Quality Standard for PM_2.5. The PM₁₀ concentrations are much higher than the WHO annual guideline 2005 mean of 20 μg/m³ and a 24-hour value of 50 μg/m³.

Source apportionments results obtained from previous Positive Matrix Factorization (PMF) analysis on both fine and coarse fractions in SR site were obtained are given in Table 3 (Begum et al., 2011a, 2010). It was observed that the contributions of motor vehicles in fine PM fraction has decreased in 2005-2006 than the previous year, but the contribution from brick kilns increased. This decrease resulted from the conversion of about 50% of vehicles to CNG. The brick industry in Bangladesh produces 15 billion bricks per year with an annual growth of about 7-8% and contributes 1% of the GDP. It consumes 2.2 million tons of coal and 1.9 million tons of fire wood. Thus, the increased size of the brick industries has driven increases in PM emission. It is also seen that number of motor vehicles, which are believed to be a major combustion source for air pollution, have increased. However, the relative BC fraction in PM_2.2 fraction has not increased over the years (Fig. 9) possibly resulting from significant reductions in the carbonaceous emissions by vehicular related sources.